<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by srinivasagopalan on Medium]]></title>
        <description><![CDATA[Stories by srinivasagopalan on Medium]]></description>
        <link>https://medium.com/@gopal010895?source=rss-f63fcd7083f9------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*8FiOMmLYk-a2HKqY</url>
            <title>Stories by srinivasagopalan on Medium</title>
            <link>https://medium.com/@gopal010895?source=rss-f63fcd7083f9------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 04:24:41 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@gopal010895/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Linearity Vs Non-Linearity in Classifying Data]]></title>
            <link>https://medium.com/@gopal010895/linearity-vs-non-linearity-in-classifying-data-6ab308a07668?source=rss-f63fcd7083f9------2</link>
            <guid isPermaLink="false">https://medium.com/p/6ab308a07668</guid>
            <category><![CDATA[classification-algorithms]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[relu]]></category>
            <dc:creator><![CDATA[srinivasagopalan]]></dc:creator>
            <pubDate>Mon, 25 May 2020 18:59:32 GMT</pubDate>
            <atom:updated>2020-05-25T18:59:32.129Z</atom:updated>
            <content:encoded><![CDATA[<p>Hi All, This is my first article in medium.</p><p>Classifying data correctly is a very big task in today’s world,although there are numerous classifiers such as (Decision Tree,Random Forrest,SVM,Logistic Regression)which does this job. It is still difficult to capture all the particular classes.</p><p>So, Lets have an example on how this Linearity and non-linearity comes into picture.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/777/1*R4fHRCjImOBbbOWifzEquw.png" /><figcaption>Fig 1. Simple Binary classification with linear line separates the classes.</figcaption></figure><p>In Fig 1, we could see that there are 2 classes 1) red stars 2) green balls.</p><p>Now if we have a linear line that separates 2 classes ,where the one above the line is stars and the one below balls.</p><p>There are high no of chances that some of the green balls could be above the line and it would eventually result in miss classification.</p><p>Now lets see what if a Non-Linear line separates the data.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/770/1*mKev89uo758kLS-6VDz4CA.png" /><figcaption>Fig 2. Simple Binary classification with Non- linear line separates the classes.</figcaption></figure><p>In Fig 2, we could again see that there are 2 classes 1) red stars 2) green balls.</p><p>Now if we have a Non-linear line like the ‘<strong>V</strong>’ shaped line in the Fig 2, that separates 2 classes ,where the one above the line is stars and the one below balls.</p><p>Here, if we observe all the data points are classified correctly.</p><p>This Non-Linearity is supported by <strong>ReLU </strong>(Rectified Linear Unit)Activation function.</p><p>Lets get to know more about <strong>ReLU</strong>;</p><p>To train deep neural networks, an activation function is needed that looks and acts like a linear function, but is, in fact, a nonlinear function allowing complex relationships in the data to be learned.</p><p>This activation function is referred to as a <strong>rectified linear activation unit</strong>, or ReLU.</p><p>The main reason to use an Activation Function in Neural Network(NN) is to introduce Non-Linearity. And ReLU does a great job in introducing the same.</p><p>ReLU is given by:</p><p><strong>R(Z) = max{0,Z}</strong></p><p>If value of Z is 0 or lesser than 0 ;R(Z) = 0</p><p>If value of Z is greater than 0 : R(Z) = Z</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/339/1*TTpjdb-MlEhCX32Za7UlPg.jpeg" /><figcaption>Fig 3) ReLU</figcaption></figure><p>As we see from the fig.3 which depicts ReLU, the values greater than zero are linear. Yet, it is a nonlinear function as negative values are always output as zero.</p><p>Since, the function is linear for half of the input and non linear for other half, it referred to as “<strong>Hinge Function</strong>”.</p><p>So, this is how classification of data can be more accurate by using <strong>ReLU</strong> which acts to be linear but is non linear in nature.</p><h3>Thanks for Reading!</h3><p>Follow me on <strong>LinkedIn</strong> <a href="https://www.linkedin.com/in/srinivas-gopal-ab9b6546/">here</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6ab308a07668" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>